Dennis Kato, Deputy Chief of Police
Los Angeles Police Department
Police have always relied on data — whether push pins tracking crimes on a map, mug shot cards, or extracting information from multiple databases and integrating them to obtain a person’s complete criminal history. The challenge facing the Los Angeles Police Department (LAPD) was that all of this information was housed in standalone databases. Analyzing this information was slow and cumbersome. The LAPD aimed to develop a data analytics strategy that was faster and could assist in disrupting crime.
The LAPD implemented several data analytic tools to accomplish this goal. The first was a relational database search program that ingests data from a variety of sources. Prompted by each search, the program links data from these sources such as address, phone number, vehicle number, and other information about a possible suspect. The linked information appears in bubbles around the suspect’s name, with lines showing the degrees of connection. Investigators are able to query each “bubble,” enabling them to identify criminal groups and crime patterns.
Another program utilizes machine-learning-based algorithms to examine four factors of property crimes – type of crime, location, time of day and day of week – to forecast locations where the probability of future crimes is high. Officers are then deployed to these locations in hopes of deterring the criminal.
Lastly, a dashboard was developed, allowing managers to compare current crime trends to past crime trends and patterns, which enables managers to compare deployment information with crime patterns to measure effectiveness.
The overall impact of the LAPD’s data-driven policing has been a downward trend in crime. The LAPD now uses near-real-time crime data to adjust deployment on a daily or even hourly basis. Real-time crime maps now identify hot spots where crimes tend to happen on certain times and days, focusing officers on very specific areas.